ensemble_deviance: Compute the deviance for models in an ensemble.
In rplzzz/CovMitigation: COVID-19 infection model for Virginia

Description Usage Arguments Details Value

The deviance for a model is 2(log L - log L0), where L is the likelihood value for the model and L0 is the likelihood for the saturated model (a hypothetical model that gets every observation exactly correct).

1	ensemble_deviance(filterfit)

filterfit

Filter-fit object for the ensemble. The history must be included in order to calculate the deviance (see fit_filter).

Theoretically, we should compute the expected observations by averaging the prevalence over the week corresponding to the observation, but we get a pretty good approximation to that by just taking the point estimates at the end of the week.

The input object, with a new entry deviance added.

rplzzz/CovMitigation documentation built on June 7, 2021, 8:48 a.m.